Rob Sandelin writes:
"Data is better than none, but when you get
into ecology, my learning is that extrapolating data to cause and effect is
always done with a great deal of caution. ...The scientist typically
hedges, knowing how little any one set of data actually tells us about the
complexity around us."

So true, although the amount of hedging depend on the type of scientist. As
an ecologist, I am an expert at hedging. It goes with the territory. I
have undergraduate degrees in physics and math. Physics is conceptually
difficult but the experimental results are often indisputable. One elegant,
well-designed experiment can make or break a break a theory. Ecology is the
reverse, conceptually easy, but almost impossible to design a definitive
experiment. Plant ecologists have been arguing about whether plants form
communities for (get this) nearly 100 years. They are still arguing about
it. In ecology, you talk about statistical significance, preponderance of
evidence, correlation, and use
words like "suggests" or "indicates.". Never say never or always.

Nonetheless, I like to think that ecologists in their stumbling path through
the maze of millions of interacting biological factors and abiotic
variability have learned a few things.

Mike Patterson notes that "1) Correlative does not mean causative." and
writes that the decline of Rufous Hummingbirds began before the significant
arrival of Anna's Hummingbird, and refers to the BBS data.

I don't want to get side-tracked on the Hummingbird question specifically,
but I checked the BBS data for Rufous and Anna's Hummingbird. In the Trend
Analysis section for Anna's Hummingbird, there are no data presented for
Washington, presumably because there aren't enough data for a trend
analysis. For Oregon, these are the Anna's Hummingbird numbers:

Oregon (Anna's)
1966-2002, 0.7% increase, nonsignificant (p=0.99).
1966-1979, no trend analysis (presumably not enough data?)
1980-2002, 1.2% increase, nonsignificant (p =0.98)

In the Trend Analysis section for Rufous Hummingbird, these are the numbers
for Oregon and Washington:

Oregon (Rufous):
1966-2002, 4.5% decline/year, statistically significant (p = 0.00)
1966-1979, 1.4% decline/year, not statistically significant (p = 0.61)
1980-2002, 2.8% decline/year, not statistically significant (p = 0.17)

Washington (Rufous)
1966-2002, 1.4% decline/year, statistically significant (p=0.02)
1966-1979, 1.4% increase/year, not statistically significant (p = 0.58)
1980-2002, 1.0% decline/year, not statistically significant (p = 0.38)

For Rufous, both Oregon and Washington data are given the blue (highest)
Regional Credibility measure, meaning the data are based on "at least 14
samples in the long-term, of moderate precision and moderate abundance."
For Anna's the Oregon data are given the yellow (intermediate) Regional
Credibility.

Both the Oregon and Washington trend analyses show the steepest Rufous
Hummingbird decline occurring since 1980. In Washington, Rufous was even
showing a (nonsignificant) increase between 1966 and 1979. The first
record of a breeding Anna's in Washington was in 1976 and they have been
increasing ever since.

Rufous Hummingbirds have been increasing since 1966 in Alaska, Alberta,
California, Idaho, and Montana. They have been decreasing since 1966 in
British Columbia, Oregon, and Washington, with the largest decreases since
1980. Rufous has decreased in the areas where Anna's has expanded its
range, but has been increasing where Anna's is either absent or has
historically been present.

Mike, how are those data inconsistent with possible competition between
Anna's and Rufous? Am I missing something?

As Mike notes, CORRELATIVE IS NOT CAUSATIVE. (It's worth putting in all
caps.) As I said in a previous post, I do not know whether Anna's and
Rufous are competing. The BBS numbers (at least by my interpretation)
suggest competition, but may represent normal fluctuations in populations,
may be an artifact of the BBS sampling methods, or may be completely
coincidental. Hummingbirds are poorly sampled by BBS techniques, so the
numbers are suspect regardless.

My point is that we don't know much about the effects of feeding. I used
Anna's as an example because it's a big favorite of backyard birders, but no
one seems to be giving much thought to whether it might be having a
detrimental impact on other species. As other posters have noted, Anna's is
closely associated with developed areas and it may benefiting as much from
the heat of cities and horticultural flowers as from feeders.

Kelly Cassidy